| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
|
| |
cleanup: Move function limit_krb5_enctypes() from the section
containing static functions into the section containing
externally visible functions.
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently if a krb5 context expires, GSSAPI authenticated RPC calls
start returning error (-EACCES in particular). This is bad when someone
has a long running job that's doing filesystem ops on a krb5 authenticated
NFS mount and just happens to forget to redo a 'kinit' in time.
The existing gssd always does a downcall with a '-1' error code if there
are problems, and the kernel always ignores this error code. Begin to
fix this by having gssd distinguish between someone that has no
credcache at all, and someone who has an expired one. In the case where
there is an existing credcache, have gssd downcall with an error code of
-EKEYEXPIRED. If there's not a credcache, then downcall with an error of
-EACCES.
We can then have the kernel use this error code to handle these
situations differently.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Recent kernels (2.6.32) have started displaying the scopeid for some
addresses in the upcall. gssd doesn't know how to deal with them. Change
gssd to use getaddrinfo instead of inet_pton since that can deal with
scopeid's in addresses. That also allows us to elminate the port
conversion in read_service_info.
If getaddrinfo returns an address with a non-zero sin6_scope_id however,
reject it. getnameinfo ignores that field and just uses the sin6_addr
part when resolving. But, two addresses that differ only in
sin6_scope_id could refer to completely different hosts.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add processing of the "service=" attribute in the new gssd upcall.
If "service" is specified, then the kernel is indicating that
we must use machine credentials for this request. (Regardless
of the uid value or the setting of root_uses_machine_creds.)
If the service value is "*", then any service name can be used.
Otherwise, it specifies the service name that should be used.
(For now, the values of service will only be "*" or "nfs".)
Restricting gssd to use "nfs" service name is needed for when
the NFS server is doing a callback to the NFS client. In this
case, the NFS server has to authenticate itself as "nfs" --
even if there are other service keys such as "host" or "root"
in the keytab.
Another case when the kernel may specify the service attribute
is when gssd is being asked to create the context for a
SETCLIENT_ID operation. In this case, machine credentials
must be used for the authentication. However, the service name
used for this case is not important.
Signed-off-by: Olga Kornievskaia <aglo@citi.umich.edu>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add processing of the "target=" attribute in the new gssd upcall.
Information in this field is used to construct the gss service name
of the server for which gssd will create a context .
This, along with the next patch handling "service=", is needed
for callback security.
For Kerberos, the NFS client will use a service principal present
in its keytab during authentication of the SETCLIENT_ID operation.
When establishing the context for the callback, the gssd on the
NFS server will attempt to authenticate the callback against the
principal name used by the client.
Note: An NFS client machine must have a keytab for the callback
authentication to succeed.
Signed-off-by: Olga Kornievskaia <aglo@citi.umich.edu>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add support for handling the new client-side upcall. The kernel,
beginning with 2.6.29, will attempt to use a new pipe, "gssd",
which can be used for upcalls for all gss mechanisms.
The new upcall is text-based with an <attribute>=<value> format.
Attribute/value pairs are separated by a space, and terminated
with a new-line character.
The intial version has two required attributes,
mech=<gss_mechanism_name> and uid=<user's_UID_number>, and two
optional attributes, target=<gss_target_name> and service=<value>.
Future kernels may add new attribute/value pairs.
Signed-off-by: Olga Kornievskaia <aglo@citi.umich.edu>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
For convenience, add the full name of the upcall pipe being processed.
(Distinquishes between "normal" upcall, and a callback upcall.)
Signed-off-by: Olga Kornievskaia <aglo@citi.umich.edu>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Change the processing so that all subdirectories within the rpc_pipefs
directory are treated equally. Any "clnt" directories that show up
within any of them are processed. (As suggested by Bruce Fields.)
Note that the callback authentication will create a new "nfs4d_cb"
subdirectory. Only new kernels (2.6.29) will create this new directory.
(The need for this directory will go away with NFSv4.1 where the
callback can be done on the same connection as the fore-channel.)
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Split out the processing for a pipe to a separate routine. The next
patch adds a new pipe to be processed.
Signed-off-by: Olga Kornievskaia <aglo@citi.umich.edu>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
nfs client used to authenticate, to the svcgssd downcall
information. This information is needed for the callback
authentication.
When estabishing the callback, nfsd will pass the principal
name in the upcall to the gssd. gssd will acquire a service
ticket for the specified principal name.
Signed-off-by: Olga Kornievskaia <aglo@citi.umich.edu>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We have a problem with rpc.gssd which blindly caches machine credentials.
E.g., if someone deletes /tmp/krb5cc_machine_REALM, rpc.gss does not create
new one until the old one expires. Also, it has problems with clock skew, if
time goes back and gssd thinks that machine credentials are not expired yet.
The following patch tries to use cache but in case of failure, it tries it
again without cache. Any comments?
Signed-off-by: Lukas Hejtmanek <xhejtman@ics.muni.cz>
Acked-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
idmapd and svcgssd have a mydaemon() routine that uses closeall() to
close file descriptors. Unfortunately, they aren't using it correctly
and it ends up closing the pipe that the child process uses to talk to
its parent.
Fix this by not using closeall() in this routine and instead, just close
the file descriptors that we know need to be closed. If /dev/null can't
be opened for some reason, then just have the child exit with a non-zero
error.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
libnfsidmapd libraries when verbosity level is set
by the '-v' flag it on either daemon.
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
All of the pieces to handle IPv6 are now in place. Add IPv6-specific
code wrapped in the proper #ifdef's so that IPv6 support works when
it's enabled at build-time.
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
We already have a common function for setting up an RPC client. That
function uses the tirpc API when tirpc is enabled and is also already
IPv6 enabled. Switch gssd to use it.
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We already have the server's address from the upcall, so we don't really
need to look it up again, and querying the local services DB for the
port that the remote server is listening on is just plain wrong.
Use rpcbind to set the port for the program and version that we were
given in the upcall. The exception here is NFSv4. Since NFSv4 mounts
are supposed to use a well-defined port then skip the rpcbind query
for that and just set the port to the standard one (2049).
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The current upcall could be more efficient. We first convert the address
to a hostname, and then later when we set up the RPC client, we do a
hostname lookup to convert it back to an address.
Begin to change this by keeping the address in the clnt_info that we get
out of the upcall. Since a sockaddr has a port field, we can also
eliminate the port from the clnt_info.
Finally, switch to getnameinfo() instead of gethostbyaddr(). We'll need
to use that call anyway when we add support for IPv6.
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If qword_eol() fails while writing the context information, log
an indication of the failure.
This addresses at least one cause of the intermittent, and
previously undiagnosed, problem of the server returning
GSS_S_NO_CONTEXT when a context was seemingly successfully
created and sent down to the kernel. In my case there was a
mis-match between kernel and user-land configuration resulting in
the proper kernel module not being loaded. Therefore the write
of the context failed, but was not logged by svcgssd. When the
kernel goes to find the resulting context, it was really not
there and correctly returned GSS_S_NO_CONTEXT to the client.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
flag has been set. This cause warnings to be generated when
return values from reads/writes (and other calls) are not
checked. The patch address those warnings.
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
this. Later code changes may make it more likely for this problem to
occur.
Also eliminate some unneeded NULL pointer checks before freeing memory.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There's no way for the caller of gssd_k5_err_msg to know whether to free
the string it returns. It can call krb5_get_error_message which returns
a string that must be freed via krb5_free_error_string. The other ways
that it can return a string require that the memory not be freed.
Deal with this by copying the string to a new buffer in all cases. Then
we can properly free the string allocated by krb5_get_error_message.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Valgrind complains that we're passing an unintialized buffer to sscanf
here. The main problem seems to be that we're not ensuring that the
buffer is NULL terminated before we pass it off.
This is the second version of this patch, the first one did not increase
the buffer allocation by 1 which could have led to clobbering the next
byte on the stack if nbytes == INFOBUFLEN.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
This causes a compiler warning and also means that we're stuffing
the buffer with uninitialized junk from the stack. Other places in
this code initialize "fakeseed" to 0. Do the same here.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
From: Jeff Layton <jlayton@redhat.com>
Make the pkgconfig check for libgssglue conditional on tirpc being
enabled. When it's disabled, the pkgconfig check for librpcsecgss will
pull in the gssglue lib and include dir automatically.
Also, make sure we include GSSGLUE_CFLAGS and the GSSGLUE_LIBS to the
appropriate places in utils/gssd/Makefile.am so that we pick up
the gssglue libs when tirpc is enabled.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Data type incompatibilities between the legacy RPC headers and the
TI-RPC headers mean we can't use libtirpc with code that was compiled
against the legacy RPC headers. The definition of rpcprog_t for
example is "unsigned long" in the legacy library, but it's "uint32_t"
for TI-RPC. On 32-bit systems, these types happen to have the same
width, but on 64-bit systems they don't, making more complex data
structures that use these types in fields ABI incompatible.
Adopt a new strategy to deal with this issue. When --enable-tirpc is
set, append "-I/usr/include/tirpc" to the compilation steps. This
should cause the compiler to grab the tirpc/ headers instead of the
legacy headers. Now, for TI-RPC builds, the TI-RPC legacy functions
and the TI-RPC headers will be used. On legacy systems, the legacy
headers and legacy glibc RPC implementation will be used.
A new ./configure option is introduced to allow system integrators to
use TI-RPC headers in some other location than /usr/include/tirpc.
/usr/include/tirpc remains the default setting for this new option.
The gssd implementation presents a few challenges, but it turns out
the gssglue library is similar to the auth_gss pieces of TI-RPC. To
avoid similar header incompatibility issues, gssd now uses libtirpc
instead of libgssglue if --enable-tirpc is specified. There may be
other issues to tackle with gssd, but for now, we just make sure it
builds with --enable-tirpc.
Note also: svc_getcaller() is a macro in both cases that points to
a sockaddr field in the svc_req structure. The legacy version points
to a sockaddr_in type field, but the TI-RPC version points to a
sockaddr_in6 type field.
rpc.mountd unconditionally casts the result of svc_getcaller() to a
sockaddr_in *. This should be OK for TI-RPC as well, since rpc.mountd
still uses legacy RPC calls (provided by glibc, or emulated by TI-RPC)
to set up its listeners, and therefore rpc.mountd callers will always
be from AF_INET addresses for now.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In gssd/context_lucid.c, ensure that gss_buffer_desc and gss_OID_desc
are defined before write_bytes.h, which uses these definitions, is
included. With TI-RPC, these definitions are not provided by
rpc/rpc.h.
It appears that <gssapi/gssapi_krb5.h> already includes krb5.h and
gssapi.h (on my system, anyway) so let's drop those includes.
Ideally write_bytes.h itself should include the needed headers, but
some source files that use Heimdal include a different, Heimdal-
compatible, header to get these definitions.
Pointed-out-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
Clean up.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Change the priority of "common" log messages so that syslog doesn't get
slammed/spammed when users' credentials expire, or there is another
common
problem which would cause error messages for all context creation
requests.
Note that this will now require that gssd or svcgssd option "-v" is used
to
debug these common cases.
Original patch from Andrew Pollock <apollock@google.com>.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
CC: Andrew Pollock <apollock@google.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of sending down an infinite expiration value for the rsi(init) and
rsc(context) cache entries, use a reasonable value for the rsi cache, and
the actual context expiration value for the rsc cache.
Prompted by a proposal from Neil Brown as a result of a complaint of a
server running out of kernel memory when under heavy load of rpcsec_gss
traffic. Neil's original patch used one minute for the init cache and one
hour for the context cache. Using the actual expiration time prevents
unnecessary context re-negotiation.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Add some plumbing so that the context expiration can be returned while
serializing the information. Later patch(es) will actually get the
expiration and pass it down to the kernel.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I have a situation where rpc.gssd appears to not be working.
Mount attempts which need to communicate with it block.
I've narrowed down the problem to that fact that all realtime signals
have been blocked. This means that DNOTIFY_SIGNAL (which is a
realtime signal) is never delivered, so gssd never rescans the
rpc_pipe/nfs directory.
It seems start_kde (or whatever it is called) and all descendants have
these
signals blocked. xfce seems to do the same thing. gnome doesn't.
So if you start rpc.gssd from a terminal window while logged in via
KDE, it doesn't behave as expected.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
environment this may not be the desired behaviour. Therefore a new
option, -R preferred realm, is presented so that the rpc.gssd prefers tickets
from this realm. By default, the default realm is preferred.
Signed-off-by: Lukas Hejtmanek <xhejtman@ics.muni.cz>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
of the Kerberos ticket used in its creation. (For contexts
created using the Kerberos mechanism.) Thus kdestroy has
no effect in nullifying the kernel context.
This patch adds -t <timeout> option to rpc.gssd so that the client's
administrator may specify a timeout for expiration of contexts in kernel.
After this timeout, rpc.gssd is consulted to create a new context.
By default, timeout is 0 (i.e., no timeout at all) which follows the
previous behavior.
Signed-off-by: Lukas Hejtmanek <xhejtman@ics.muni.cz>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
found. Trying to use invalid default cache and continue is not good idea at all.
Signed-off-by: Lukas Hejtmanek <xhejtman@ics.muni.cz>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
possible to search several directories for valid credentials when
making NFS requests.
Original patch from Vince Busam <vbusam@google.com>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>.
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
so that some messages that would otherwise always print may
be silenced.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
Kerberos encryption types that may be negotiated.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
see if a port number was supplied. If so, use it rather
than the default port number.
Signed-off-by: Olga Kornievskaia <aglo@citi.umich.edu>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
| |
Modified-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
| |
Signed-off-by: Steinar H. Gunderson <sgunderson@bigfoot.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
the size of the poll array
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
defined by the -d flag
Signed-off-by: Steve Dickson <steved@redhat.com>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
use the new xlog logging infrastructure.
This patch removes all of the old idmap_* logging functions and replaced them
with the corresponding xlog functions. In addition that that it also reworks
the gssd logging wrappers to use the new xlog_backend. Finally it makes
necessary changes to the build files to get the project compiling again.
Signed-off-by: David P. Quigley <dpquigl@tycho.nsa.gov>
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
|
|
|
|
|
|
|
|
| |
Now that svcgssd is using the qword_* functions in the
support library, remove the private version.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
|
|
|
|
|
|
|
|
|
|
| |
Now that the nfslib library has all the necessary functions and they
all operate as needed, use them instead of the private versions in
utils/gssd/cacheio.c.
The obsolete private versions are removed in the next patch.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
|
|
|
|
|
|
|
|
|
|
|
| |
Rather than depending on modified qword_* functions to print
svcgssd debugging information, use printerr in the downcall
function.
And while we're at it, label things so we know what we're looking at!
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
|
|
|
|
|
|
|
|
|
|
| |
print_hexl() currently uses printerr, but is really only necessary
for local debugging and should simply write to stdout.
Also change it to print the description internally.
Wrap it and its use in #ifdef DEBUG.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
|
|
|
|
|
|
|
|
| |
Fix function declaration to eliminate compiler warning about it
not being a prototype after -Wstrict-prototypes was added.
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
|
|
|
|
|
|
|
| |
Fix the usage message for gssd to reflect new -M option added in 1.1.0
Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
|
|
|
|
|
|
|
| |
Update gitignore to ignore some generated files.
Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
|